Recent advances in deep learning for speech research at Microsoft

TLDR

Deep learning is becoming a mainstream technology for speech recognition at industrial scale. This paper reviews Microsoft speech research since 2009, highlighting recent advances that illuminate the core capabilities and limitations of current deep‑learning methods. The authors organize the review by feature‑domain and model‑domain dimensions, presenting experimental results from speech recognition, spoken dialogue, and language modeling to illustrate the strengths and weaknesses of the techniques. The study discusses the experimental findings, identifies potential improvements, and outlines future research directions for these methods.

Abstract

Deep learning is becoming a mainstream technology for speech recognition at industrial scale. In this paper, we provide an overview of the work by Microsoft speech researchers since 2009 in this area, focusing on more recent advances which shed light to the basic capabilities and limitations of the current deep learning technology. We organize this overview along the feature-domain and model-domain dimensions according to the conventional approach to analyzing speech systems. Selected experimental results, including speech recognition and related applications such as spoken dialogue and language modeling, are presented to demonstrate and analyze the strengths and weaknesses of the techniques described in the paper. Potential improvement of these techniques and future research directions are discussed.